The Highly Regioselective Synthesis of Novel Imidazolidin-2-Ones via the Intramolecular Cyclization/Electrophilic Substitution of Urea Derivatives and the Evaluation of Their Anticancer Activity

A series of novel 4-(het)arylimidazoldin-2-ones were obtained by the acid-catalyzed reaction of (2,2-diethoxyethyl)ureas with aromatic and heterocyclic C-nucleophiles. The proposed approach to substituted imidazolidinones benefits from excellent regioselectivity, readily available starting materials and a simple procedure. The regioselectivity of the reaction was rationalized by quantum chemistry calculations and control experiments. The anti-cancer activity of the obtained compounds was tested in vitro.


Introduction
Various cyclic ureas are the well-established pivotal functionalities in the drug discovery and medicinal chemistry. Among them, the imidazolidine-2-one scaffold is often found in bioactive compounds and constitute the key structural part of a plethora of FDA approved drugs, such as emicerfont, imidapril, azlocillin and others ( Figure 1). Moreover, these compounds may serve as precursors for vicinal diamines, which are valuable building blocks in organic and medicinal chemistry [1,2]. Taking these into account, it is not surprising that a lot of effort is devoted to the development of the methods of synthesis of imidazolidinones [3].
The most straightforward approach to cyclic ureas is the carbonylation of variously substituted diamines. This approach is somewhat controversial due to the abovementioned utility of imidazolidine-2-ones as 1,2-diamines precursors. However, it is still used in some cases, especially with chiral diamines [4][5][6][7]. Novel approaches to these compounds have been developed in 2019-2020 by various research groups, indicating a growing area of interest. These include silver-catalyzed cycloaddition of nitrones with methylene isocyanides [8], stereoselective diamination of alkenes with 1,3-ditosylurea [9] and Rh-catalyzed intramolecular C-H amination of N-oxyurea derivatives [10,11]. Pdcatalyzed amidation of vinylethylcarbamates followed by cyclization of intermediate Rh-catalyzed intramolecular C-H amination of N-oxyurea derivatives [10,11]. Pd-catalyzed amidation of vinylethylcarbamates followed by cyclization of intermediate allyl ureas should also be mentioned [12]. Finally, the amidoalkylation of indoles by hydroxyimidazolidin-2-one under the Lewis acid catalysis has also been reported [13]. Nitrogen-containing aldehydes and acetals are widely used in the heterocycles synthesis [14,15]. Earlier, we have described the approach to 2-substituted pyrrolidine derivatives based on the in situ generation of a cyclic iminium ion from 4-aminobutanal acetals (4,4-diethoxybutan-1-amines) and its trapping by various nucleophiles [16][17][18]. Our preliminary studies indicate that N- (2,2-dialkoxyethyl) ureas are also capable of the formation of cyclic imidazolinium cations and the reaction results in cyclic urea derivativesimidazolidinones or benzo [d] [1,3] diazepinones [19]. As the continuation of our efforts, herein, we report the successful application of this methodology to the synthesis of the series of 4-(hetero) arylimidazolidin-2-ones starting from N- (2,2-dialkoxyethyl) ureas and (hetero) aromatic nucleophiles (Scheme 1). The main advantages of the described method are the mild reaction conditions, good to high product yield and the high regioselectivity. Notably, N- (2,2-dialkoxyethyl) ureas are easily accessible from commercially available 2,2-dimethoxyethan-1-amine and amines or isocyanates in just one step, which also adds value to the proposed approach. Some considerations on the regioselectivity of the reaction are discussed based on quantum chemistry calculations, as well as evaluation of the in vitro anti-cancer activities of the obtained compounds. Scheme 1. The proposed synthetic approach to 4-(hetero) arylimidazolidine-2-ones.

Chemistry
We initiated our studies with the expanding the scope of the starting N-(2,2-dialkoxyethyl) ureas 1 (Scheme 2). Di-and trisubstituted ureas possessing various aryl (1a-f) and aryl/alkyl (1g-m) groups were easily obtained from appropriate isocyanates by known procedures [20]. Nitrogen-containing aldehydes and acetals are widely used in the heterocycles synthesis [14,15]. Earlier, we have described the approach to 2-substituted pyrrolidine derivatives based on the in situ generation of a cyclic iminium ion from 4-aminobutanal acetals (4,4diethoxybutan-1-amines) and its trapping by various nucleophiles [16][17][18]. Our preliminary studies indicate that N-(2,2-dialkoxyethyl) ureas are also capable of the formation of cyclic imidazolinium cations and the reaction results in cyclic urea derivatives-imidazolidinones or benzo [d] [1,3] diazepinones [19]. As the continuation of our efforts, herein, we report the successful application of this methodology to the synthesis of the series of 4-(hetero) arylimidazolidin-2-ones starting from N-(2,2-dialkoxyethyl) ureas and (hetero) aromatic nucleophiles (Scheme 1). The main advantages of the described method are the mild reaction conditions, good to high product yield and the high regioselectivity. Notably, N-(2,2dialkoxyethyl) ureas are easily accessible from commercially available 2,2-dimethoxyethan-1-amine and amines or isocyanates in just one step, which also adds value to the proposed approach. Some considerations on the regioselectivity of the reaction are discussed based on quantum chemistry calculations, as well as evaluation of the in vitro anti-cancer activities of the obtained compounds. Rh-catalyzed intramolecular C-H amination of N-oxyurea derivatives [10,11]. Pd-catalyzed amidation of vinylethylcarbamates followed by cyclization of intermediate allyl ureas should also be mentioned [12]. Finally, the amidoalkylation of indoles by hydroxyimidazolidin-2-one under the Lewis acid catalysis has also been reported [13]. Nitrogen-containing aldehydes and acetals are widely used in the heterocycles synthesis [14,15]. Earlier, we have described the approach to 2-substituted pyrrolidine derivatives based on the in situ generation of a cyclic iminium ion from 4-aminobutanal acetals (4,4-diethoxybutan-1-amines) and its trapping by various nucleophiles [16][17][18]. Our preliminary studies indicate that N- (2,2-dialkoxyethyl) ureas are also capable of the formation of cyclic imidazolinium cations and the reaction results in cyclic urea derivativesimidazolidinones or benzo [d] [1,3] diazepinones [19]. As the continuation of our efforts, herein, we report the successful application of this methodology to the synthesis of the series of 4-(hetero) arylimidazolidin-2-ones starting from N- (2,2-dialkoxyethyl) ureas and (hetero) aromatic nucleophiles (Scheme 1). The main advantages of the described method are the mild reaction conditions, good to high product yield and the high regioselectivity. Notably, N- (2,2-dialkoxyethyl) ureas are easily accessible from commercially available 2,2-dimethoxyethan-1-amine and amines or isocyanates in just one step, which also adds value to the proposed approach. Some considerations on the regioselectivity of the reaction are discussed based on quantum chemistry calculations, as well as evaluation of the in vitro anti-cancer activities of the obtained compounds. Scheme 1. The proposed synthetic approach to 4-(hetero) arylimidazolidine-2-ones.

Chemistry
We initiated our studies with the expanding the scope of the starting N-(2,2-dialkoxyethyl) ureas 1 (Scheme 2). Di-and trisubstituted ureas possessing various aryl (1a-f) and aryl/alkyl (1g-m) groups were easily obtained from appropriate isocyanates by known procedures [20]. Next, we investigated the reaction of the ureas 1a-m with various electron-rich aromatic and heterocyclic C-nucleophiles (Scheme 3). In all cases, the reaction provided the desired imidazolidinones 2 in good to high yield. The previously reported conditions Scheme 2. The synthesis of starting N-(2,2-dialkoxyethyl)ureas 1.
Next, we investigated the reaction of the ureas 1a-m with various electron-rich aromatic and heterocyclic C-nucleophiles (Scheme 3). In all cases, the reaction provided the desired imidazolidinones 2 in good to high yield. The previously reported conditions were employed, i.e., refluxing toluene and a triflouroacetic acid (TFA) as catalysts [19]. Notably, the 15-fold excess of catalyst was used in these preliminary experiments. However, the further screening of reaction conditions allowed us to decrease the amount of catalyst significantly without a loss of yield. Moreover, during our preliminary studies, we found that this reaction leads to the mixture of 4-and 5-substituted imidazolidin-2-ones, with 4-substituted regioisomer 2 being the major one. Decreasing the amount of TFA led to the improved regioselectivity, so that only 4-substituted imidazolinones 2 were observed in the reaction mixture. The isolated yields of target compounds varied somewhat. In general, ureas possessing unsubstituted nitrogen atom tend to provide higher yields of imidazolidinones 2. However, no correlation was observed between the yields and the substituents in the aryl moiety of the starting ureas.   The chemical shifts of methyne and methylene protons for 4-and 5-substituted imidazolidin-2-ones differ notably [19] (see Supplementary Materials, Figure S2). Thus, the substitution site of the imidazolidinone ring for the N-methyl substituted compounds 2h-m and 4h-l was determined by the comparison of their NMR spectra with the spectra of previously described 4-substituted imidazolidin-2-ones, 2g and 4g [19] (for example, see Supplementary Materials, Figures S3 and S4). Additionally, the structures of the compounds 2j and 2k were confirmed by X-ray analysis ( Figure 2; for details, see Supplementary Materials, pp. S2-S3 and Figure S1).    Interestingly, the reaction of the urea 1d with 4-hydroxy-6-methylpyran-2-one re sulted in the mixture of regioisomeric 4-(2-oxopyran-3-yl)-and 4-(2-oxopyran-5-yl)imid azolidinones 5d and 5d′ (~10:1 molar ratio according to NMR data). The structure of reg oisomers was unequivocally determined by 1 H-13 C HMBC experiments ( Figure 4). Thu in compound 5d, a cross-peak is observed between the protons of the methyl group of th  Interestingly, the reaction of the urea 1d with 4-hydroxy-6-methylpyran-2-one resulted in the mixture of regioisomeric 4-(2-oxopyran-3-yl)-and 4-(2-oxopyran-5-yl)imidazolidinones 5d and 5d (~10:1 molar ratio according to NMR data). The structure of regioisomers was unequivocally determined by 1 H-13 C HMBC experiments ( Figure 4). Thus, in compound 5d, a cross-peak is observed between the protons of the methyl group of the pyran-2-one fragment and 5 -C carbon atom, which is bonded to hydrogen. At the same time, in the compound 5d , the cross-peak between the protons of the same methyl group and the substituted 5 -C carbon atom is present (see Supplementary Materials, Figure S98, for the HMBC spectrum). Additionally, there are cross-peaks between the 5 -C carbon atom and protons of the methyne (4-CH) and methylene (5-CH 2 ) groups of the imidazolidin-2-one ring. Interestingly, the reaction of the urea 1d with 4-hydroxy-6-methylpyran-2-one resulted in the mixture of regioisomeric 4-(2-oxopyran-3-yl)-and 4-(2-oxopyran-5-yl)imidazolidinones 5d and 5d′ (~10:1 molar ratio according to NMR data). The structure of regioisomers was unequivocally determined by 1 H-13 C HMBC experiments ( Figure 4). Thus, in compound 5d, a cross-peak is observed between the protons of the methyl group of the pyran-2-one fragment and 5′-C carbon atom, which is bonded to hydrogen. At the same time, in the compound 5d′, the cross-peak between the protons of the same methyl group and the substituted 5′-C carbon atom is present (see Supplementary Materials, Figure S98, for the HMBC spectrum). Additionally, there are cross-peaks between the 5′-C carbon atom and protons of the methyne (4-CH) and methylene (5-CH2) groups of the imidazolidin-2-one ring. The reaction of urea 1a with 2-methylesorcinol possessing two reactive sites resulted in the disubstituted imidazolinone 8. The compound 8 was obtained as the mixture of the (R,S)-and (S,S)-, (R,R)-diastereomers (dr = 1:1 according to NMR data), from which one diastereomer was isolated in the individual form in 16% yield. Unfortunately, we were not able to obtain the crystals suitable for the X-ray analysis; thus, the absolute configuration of the isolated compound could not be determined. The reaction of urea 1a with 2-methylesorcinol possessing two reactive sites resulted in the disubstituted imidazolinone 8. The compound 8 was obtained as the mixture of the (R,S)-and (S,S)-, (R,R)-diastereomers (dr = 1:1 according to NMR data), from which one diastereomer was isolated in the individual form in 16% yield. Unfortunately, we were not able to obtain the crystals suitable for the X-ray analysis; thus, the absolute configuration of the isolated compound could not be determined.

Mechanism Discussion and Quantum Chemistry Calculations
The proposed mechanism of the formation of imidazolidine-2-ones from ureas 1 is depicted in Scheme 4 using phenol as the model nucleophile. The first stage of the reaction is the formation of the oxonium cation and its subsequent intramolecular cyclization to Quantum chemistry calculations were performed for iminium ions B and D, as well as for O-protonated ion E to justify the regioselectivity of the reaction. The relative energies of the intermediates are given in Scheme 4. The oxonium ion E appeared to be the most stable, which is in accordance with the preferred protonation of the ureas at the oxygen atom [21]. The energy of the iminium ion D is 7.3 kcal/mol higher compared to cation E. Surprisingly, the isomeric iminium ion B appeared to be somewhat lower in energy (6.8 kcal/mol).
Next, energies of the intermediates F and G, as well as energies of the appropriate transition states TS2 and TS3, were calculated. ortho-Substitution of the phenol was modelled to reflect the presence of the hydroxy group next to the reactive center in the most of the used nucleophiles. Again, the relative energy of the intermediate F was lower than that of the intermediate G. At the same time, the relative energy of the transition state TS3 appeared to be lower than the energy of transition state TS2 by circa 5 kcal/mol. According to the Curtin-Hammett principle [22], the product distribution for the two rapidly interconverting intermediates depends both on their relative energies and energies of the respective transition states. Thus, the observed regioselective formation of 4-substituted imidazolidine-2-ones may be attributed to its lower activation energy.
is the formation of the oxonium cation and its subsequent intramolecular cycliza give 5-methoxyimidazolidine-2-one A. The acid-promoted elimination of the m molecule furnishes iminium cation B. This intermediate may further undergo two rent reactions. The first one is the interaction with phenol molecule, which result intermediate Quantum chemistry calculations were performed for iminium ions B and D, as for O-protonated ion E to justify the regioselectivity of the reaction. The relativ gies of the intermediates are given in Scheme 4. The oxonium ion E appeared to most stable, which is in accordance with the preferred protonation of the ureas at ygen atom [21]. The energy of the iminium ion D is 7.3 kcal/mol higher compared t E. Surprisingly, the isomeric iminium ion B appeared to be somewhat lower in ene kcal/mol).
Next, energies of the intermediates F and G, as well as energies of the appr transition states TS2 and TS3, were calculated. ortho-Substitution of the phenol wa elled to reflect the presence of the hydroxy group next to the reactive center in th of the used nucleophiles. Again, the relative energy of the intermediate F was low that of the intermediate G. At the same time, the relative energy of the transition st appeared to be lower than the energy of transition state TS2 by circa 5 kcal/mol. Ac to the Curtin-Hammett principle [22], the product distribution for the two rapidl

Biological Studies
Some of the resulting compounds were tested for cytotoxicity against normal and cancer human cell lines at concentrations of 1-100 µM. As seen from the Table 1, most of the obtained compounds exhibited little to no cytotoxicity against both normal and cancer cells. Compounds 2h and 2k appeared to be the most active. The selectivity index (SI) was calculated for them as the ratio between the IC 50 value for normal cells and the IC 50 value for cancer cells. The cytotoxicity of the compound 2h against the HuTu 80 cancer cell line was almost two-fold higher than the cytoxicity against normal cells (SI = 1.7). This is better than the selectivity of the reference compound Arglabin (SI = 1.2). Notably, the replacement of the methyl substituent at the nitrogen atom by hydrogen lowers the cytotoxicity against cancer cells, whereas the cytotoxicity against normal cells tends to remain the same or even increases (compare, e.g ., 2h and 2b, 2k and 2d and 2j and 2c). Thus, it appears to be one of the key factors influencing the activity of the compounds under study.

Chemistry
IR spectra were recorded on a UR-20 spectrometer in a 400-3600 cm −1 range in KBr. 1 H NMR spectra were recorded on a Bruker MSL 400 spectrometer (400 MHz) with respect to the signals of residual protons of deuterated solvent (CDCl 3 , DMSO-d 6 ). 13 C NMR spectra were recorded on a Bruker Avance 600 (151 MHz) spectrometer relative to signals of residual protons of deuterated solvent (CDCl 3 , DMSO-d 6 ). Elemental analysis is performed on a Carlo Erba device EA 1108. The melting points are determined in glass capillaries on a Stuart SMP 10 instrument.
The X-ray diffraction data for the crystals of 2j and 2k were collected on a Bruker D8 Venture automatic diffractometer using graphite monochromated radiation. The structures were solved by direct methods and refined by full-matrix least-squares using the SHELXL97 [23] program. All the non-hydrogen atoms were refined with anisotropic atomic displacement parameters. All figures were made using the program OLEX2 [24]. Crystallographic data for the structure reported in this paper have been deposited with the Cambridge Crystallographic Data Center (2068964-2068965).

4-(1-phenyl-3-methyl-2-oxoimidazolidin-4-yl)-1,5-dimethyl-2-phenyl-1,2-dihydro-3H-
pyrazol-3-one (4g). Yield 60%. The spectral characteristics were in agreement with the previously published data [19]. well, and cultured in a CO 2 incubator at 37 • C. Twenty-four hours after seeding the cells into wells, the compound under study was added at a preset dilution, 150 µL to each well. The dilutions of the compounds were prepared immediately in nutrient media; 5% DMSO, which does not induce inhibition of cells at this concentration, was added for better solubility. The experiments were repeated three times. Intact cells cultured in parallel with experimental cells were used as a control.

Quantum Chemistry Calculations
All calculations have been performed with the Gaussian 16 package [29]. The initial structures were fully optimized at the B3LYP/6-311++G(d,p) theory level. All optimizations were followed by frequency calculations at the same level of theory in order to check that optimized structures really correspond to true minima.

Conclusions
In conclusion, a series of novel 4-(het) arylimidazolidines were obtained via trifluoroacetic acid catalyzed intramolecular cyclization/Mannich-type reaction of N-(2,2diethoxyethyl) ureas with high regioselectivities. The plausible mechanism was discussed using quantum chemistry calculations and the choice of the solvent and catalyst amount was found to have a crucial effect on the regioselectivity. The anti-cancer activities of the obtained compounds were tested in vitro, and the substitution of the nitrogen atom was identified as one the key factors influencing the cytotoxicity against normal and cancer human cell lines.